• Buffalox@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    7 hours ago

    Weird article IMO??? Forget the startups, unless they are backed by very major players they don’t stand a chance. They have their competences, but I seriously doubt competing with Nvidia on making AI chips is among them.
    The players to watch are Huawei, Baidu, Tencent and the likes. Who have already been working on this for a while, and have actual working and useful products.

    While Huawei is the leader, Chinese companies don’t want to rely entirely on the company.

    I don’t understand how that statement is supposed to make much sense? When Chinese companies were happy using Nvidia and being dependent on Nvidia. Why wouldn’t they be equally happy using Huawei if it’s the best option after the government has forbidden them from using Nvidia?
    It may be true, but there is zero explanation why it is.

    To the ones that think China can just use AMD instead, they really can’t, AMD is under the same restrictions Nvidia is, and AFAIK AMD has not designed a chip to sell to China within those restrictions.

  • survirtual@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    ·
    8 hours ago

    Software like wgpu makes it much easier to close the gap between various GPUs. New compute languages that are backend-agnostic are appearing, in the same vein as taichi-lang, that make it significantly easier to make high-performance gpu kernels deployable anywhere.

    The compute groundwork for crossplatform tensor calculations is already here. Inference is already doable on any device. Training is not far behind. As a side-effect of this, processing on the GPU in every capacity, like physics, novel rendering techniques, or whatever else the imagination can muster, is now within grasp of “average” programmers.

    If you have always been intimidated by GPU programming, I urge you to take another look now. The landscape is radically different. The software moat everyone talks about with NVIDIA is smoke-and-mirrors. Cuda is old news, though I am speaking to the actual code landscape here, not the common mental consensus.

    What we lack now is cheap video cards that have high memory. I believe the current cards are overpriced by about 10 - 100x what they should be, because this profit situation is extremely temporary. Just as pens were once thousands of dollars, these compute devices will be collapsing in price.

    I welcome China building cheaper video cards. Hopefully we will all benefit from it before any robot wars break out.

  • mal3oon@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    ·
    9 hours ago

    AMD entered the chat.

    They’ve been trying to do it for ages, and they’re forever a tad closer. The software stack is where the challenge lies, not the hardware. Still today, you can buy AMD hardware that on paper is better than Nvidia’s, yet you can’t squeeze out similar performances matching shittier Nvidia cards. Even still, rocm, the ‘cuda’ for AMD, can’t even compete with even vulkan (an open source agnostic backend). So I doubt china will deliver that fast.

    • Buffalox@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      7 hours ago

      AMD is under the same restrictions Nvidia is, but hasn’t designed a chip within those restrictions to sell to China like Nvidia did.